Case 6: Missed small bowel injury following complex revisional bariatric surgery

General surgery

CASE SUMMARY A female patient in her mid-60s was admitted to a private hospital for elective weight loss surgery. She was morbidly obese, weighing 135 kg with a body mass index of 51. Medical history included a gastric band of some 20 years, with consequent band failure, pouch dilatation and dysmotility manifested by residual food in the pouch on multiple endoscopies, severe reflux and a hiatal hernia with a past Cameron ulcer. Total colectomy for ulcerative colitis had been performed more than 20 years previously, and her midline was characterised by a large ventral hernia. She had no other significant medical problems. The five-and-a-half-hour surgery entailed on-table gastroscopy, laparoscopic adhesiolysis, explantation of the gastric band, small bowel resection, laparoscopic hiatus hernia repair, and laparoscopic sleeve gastrectomy converted intraoperatively to open single loop gastric bypass following further open adhesiolysis. A small bowel injury during the adhesiolysis resulted in a small bowel resection with a primary stapled anastomosis. The early postoperative course was notable for tachycardia, tachypnoea, poor oxygen saturations, and an inadequate and declining urine output. As the morning of postoperative day one progressed, the patient became hypotensive (down to systolic in the low 70s), which was addressed with metaraminol. Agents, including ketamine and fentanyl, were employed to manage problematic pain. She became increasingly unwell during the day. Although afebrile and no longer tachycardic, she developed oliguria, and was later intubated and ventilated because of respiratory and lactic acidosis. The reason for her deterioration was unknown; consideration was given to acute chest syndrome/acute respiratory distress syndrome/intra-abdominal sepsis. The patient remained hypotensive and oliguric during the evening, with increasing vasopressor requirements. She had very poor cardiac function on echocardiogram (ejection fraction 35%) and worsening hypoxaemia. A CT scan revealed a few small pockets of intraperitoneal air but no free fluid. Discussion between the operating surgeon and ICU specialists concluded that this was a respiratory/cardiac issue. In the early hours of postoperative day two, the patient deteriorated further and was transferred to a public hospital, where she was assessed by the surgical team.

A CT pulmonary angiogram was performed, and no pulmonary embolus found. The upper gastrointestinal surgical team reviewed the patient and diagnosed intraabdominal sepsis. She was taken to theatre for emergency laparotomy and a small bowel injury was found with significant contamination. The patient was critically unwell, so damage control laparotomy was performed, with resection of the small bowel enterotomy without anastomosis, and laparostomy. The following day, a relook showed a necrotising infection in the left abdominal wall requiring extensive debridement. The small gut was viable, and no further bile leak was seen. At relook 48 hours later, the small bowel was anastomosed. The patient returned to theatre on many occasions over the ensuing eight weeks. She developed multiple fistulae, including gastrocutaneous from the proximal pouch and enterocutaneous further down the small bowel. Despite multidisciplinary management of this complex patient, she succumbed to overwhelming sepsis almost two months later.

DISCUSSION This was a complex revisional bariatric procedure in a morbidly obese woman in her mid-60s. Unfortunately for this patient, the delay in early surgical intervention rendered her course prolonged and ultimately fatal. It is apparent from the operative report that significant dissection and adhesiolysis were required to free up the anatomy to perform the planned procedure. Of significant note, after this extensive dissection, the original planned bariatric procedure was converted to another procedure intraoperatively. Of equal note, during the course of the surgery the laparoscopic approach was converted to open surgery. Within 24 hours of the procedure, the patient was seriously unwell, such that she required intubation and mechanical ventilation. The decline was initially heralded by tachycardia, tachypnoea and oliguria. Soon after, hypotension requiring inotropic support, acute kidney injury, and finally, respiratory/lactic acidosis, declared the patient critically unwell. In the background, she had pain that was very difficult to manage, requiring opioid and ketamine infusions. In the setting of a long and complex surgery, particularly one where there is a long staple line, at least two gut anastomoses and a laborious adhesiolysis, one would have expected an abdominal focus to be given first consideration as the culprit in this patient’s serious decline. It appears from the notes that the surgeon and intensivists involved in her care gave inadequate credence to this likelihood, instead focusing on a cardiorespiratory aetiology. The relatively normal CT does not exclude a major abdominal catastrophe, given the clinical signs in this morbidly obese patient.

The opportunity to salvage this patient was probably missed in the first 24 hours following surgery. It seems likely that she already had bile peritonitis at that juncture, and a timely relook laparotomy would have identified this injury before she became even more critically unwell. Early intervention may have required only a simple oversewing of an enterotomy or perhaps resection, but, at the very least, the ongoing insult would have been contained and the physiological insult limited. A second, albeit possibly more controversial, consideration is the surgeon’s choice to convert from sleeve to single loop bypass. While the principle of conversion on the basis of creating a low-pressure system (bypass) to decompress the high-pressure system (sleeve) is supported, the option of a classic Roux-en-Y configuration would be favoured for the simple reason that if this were to leak (at the gastrostomy-jejunostomy), it is a ‘dry’ limb. There are only salivary secretions passing through the alimentary limb, as the hepato-pancreato-biliary limb (along with its far more corrosive enzymes) has been diverted away from the gastrojejunostomy, typically by a length of at least 50 cm. Thus, to dry out the alimentary limb requires only a nasogastric/nasojejunal tube. In contrast, the single loop anastomosis has 2–3 L of gastric, bile and pancreatic secretions passing by, with far worse consequences in the event of an anastomotic leak. This case was already difficult due to the adhesions; furthermore, the gastric band adhesions at the top end had compromised the sleeve, which was then converted to a gastric pouch, again favouring the option of the Roux-en-Y.

CLINICAL LESSONS In the early postoperative period, certainly within the first 24 hours, a seriously unwell surgical patient must have an abdominal cause for decline excluded, even if this must be proven by reoperation. This is particularly relevant in morbidly obese patients whose physiology and clinical signs may mask serious intra-abdominal pathology. Given the scale and complexity of surgery, it is difficult to imagine that decent drain tubes were not in place after the first operation. The output of these would have signalled the bowel leak and led to earlier intervention. A very high index of suspicion is prudent and equally a low threshold for take back is safe. A negative diagnostic laparoscopy/laparotomy is far easier to justify than a delayed diagnosis resulting in a poor outcome. The missed diagnosis of peritonitis was critical. The site of initial surgery was problematic. By the second postoperative day, the private hospital was considered inadequate and transfer to a public hospital was undertaken. Given the fairly predictable serious complications with this high-risk case, the initial choice of hospital was an issue. The operating surgeon knows the patient better than anyone else involved in their

care. Thus, the surgeon should be cautious not to be unduly influenced by other specialists involved in the patient’s care when this advice seems contrary to their surgical opinion or the patient’s clinical state. The opinions of other specialists should certainly be considered and are often very valuable, but in situations such as this, sometimes the simple answer is to go back for another look and ask: What abdominal injury has caused the decline? Have I proven without doubt that the abdomen is not the culprit? When considering revisional bariatric surgery, which carries a higher risk for serious complications, procedural choice is always best tailored to a patient’s individual characteristics and previous operations. Options for conversion are largely the surgeon’s choice, with the data supporting a multitude of options. It is worth remembering that two-stage procedures are a reasonable option. Sometimes it is better to bail and come back to fight another day.